Table 1 Overall and marginal mean of rescaled pAUC5 of SEM-BayesCΠ and SEM-GWAS based on overall,direct and indirect effect
Overall effectDirect effectIndirect effect
factorsSEMBayesCПSEM-GWASSEMBayesCПSEM-GWASSEMBayesCПSEM-GWAS
trait 1nQTL
304.994.664.994.66NANA
902.001.912.001.91NANA
3001.381.421.381.42NANA
shape (γ)
0.182.712.522.712.52NANA
1.482.932.792.932.79NANA
3.002.722.692.722.69NANA
overall2.792.662.792.66NANA
trait 2nQTL
305.494.744.223.914.883.15
902.092.081.871.871.911.72
3001.341.381.301.331.281.30
shape (γ)
0.183.102.802.482.362.792.29
1.483.032.842.562.382.632.46
3.002.842.602.362.382.642.43
overall2.992.742.462.372.692.39
  • For both trait 1 and trait 2, comparisons between methods are made for different number of QTL (nQTL) and shape parameters of QTL effects (γ). Estimations are based on 450 simulated data sets including nine scenarios discussed in Simulated data. For each effect, in each row, the values with different symbols have significantly different (p-value <0.1) pAUC5. NA represented pAUC5 was not available because the indirect effect on trait 1 does not exist based on the causal structure. The overall effect result on trait 1 is the same as the direct effect result on trait 1 because the overall effect on trait 1 equals the direct effect on trait 1 based on the causal structure.